Gesture Recognition for Remote Control Systems Using Kernel Principal Component Analysis (KPCA) and Trajectory Normalization
Raghab Singh · Lecture notes in networks and systems · 2026
The paper integrates Kernel Principal Component Analysis (KPCA) along with trajectory normalization for effective classification of hand gestures employed in non-contact remote control systems. Data from the mouse tracking and vision system is supported source-agnostic recognition. Trajectory normalization helps to equalize variations in input gestures influenced by the gesture speed, orientation, and sampling imperfections, while KPCA captures nonlinear gestures in lower dimensions. The framework includes and analyzes different interpolation methods to better align execution times between inputs. The experimental results show better accuracy in recognizing gestures that are complex and degenerate, particularly when the user is inconsistent and the input is noisy. Moreover, the paper gives a thorough examination of the trade-offs of different normalization schemes and their effect on classifier performance. The development of intuitive and reliable human-computer interaction systems for smart environments and assistive technologies is contributed by this work. The experiments indicate that, applying distance-based normalization with a uniform scaling factor, the system achieves over 99.9% accuracy for non-degenerate gestures and over 98% accuracy for degenerate gestures. The system generalizes well to different input types. The study would bring real-world contactless control applications for assistive and industrial domains.